Estimation of Cognitive Load of a Pilot Under Time and Task Pressure

Abstract

Background: The growing integration of manned and unmanned aerial vehicles (MAVs and UAVs) in modern high-tech warfare demands a thorough understanding of pilot cognitive performance. Current battlefield decision-making models treat pilots as infallible operators, neglecting the impact of complex operational environments on their cognitive decision-making abilities, leading to a gap between computational outcomes and real combat scenarios.

Methods: This study proposes a multi-modal pilot cognitive decision-making model grounded in information theory. An interference characteristic matrix was constructed to quantify the influence of cooperative combat characteristics (UAV count, flight tasks, communication link quality), battlefield environmental factors (visibility level, task urgency), cockpit human-machine interface efficiency, and pilot personal capabilities on cognitive performance. A semi-physical simulation experiment was conducted with 27 participants using a Boeing 777-300 cockpit simulator equipped with an SMI iView X HED helmet-mounted eye-tracker. A within-subject repeated-measures design encompassed 54 task conditions across varying visibility levels (CAVOK, low visibility, night navigation), UAV counts (1, 2, 3), communication link quality (good, poor), and flight tasks (landing, climbing, level flight). Dependent variables included visual entropy (composite eye tracking indicator), reaction time (RT), accuracy (ACC), and subjective cognitive load assessed via NASA-TLX and 3D-SART scales.

Results: Statistical analysis confirmed significant effects of all independent variables on cognitive performance. Cognitive decision-making levels decreased with increasing UAV count, reflecting heightened information load. Night navigation conditions yielded higher cognitive decision making levels due to increased display contrast. Good communication link quality facilitated higher cognitive performance. Landing tasks imposed lower cognitive decision-making levels than climbing or level flight. The proposed multimodal model demonstrated strong correlation with empirical results across all 54 task conditions, outperforming the classical single-indicator 3D-SART method.

Conclusions: The interference characteristic matrix successfully quantifies the influential factors of pilot cognition and decision-making in MAV/UAV cooperative engagements. The multimodal model, integrating visual entropy, performance metrics, and subjective cognitive load, provides a comprehensive and accurate characterization of pilot cognitive decision-making ability. These findings offer significant practical implications for cockpit human-machine interface (HMI) design, pilot combat training optimization, and aviation safety improvement.

Keywords: Manned Aerial Vehicles (MAV); Unmanned Aerial Vehicles (UAV); Pilot Cognitive Load; Visual Entropy; Eye-Tracking; Multimodal Decision Model; Cooperative Combat; Information Theory

Citation

Kalam FA, Sultan M, Zhenbao L. Estimation of Cognitive Load of a Pilot Under Time and Task Pressure. WebLog J Comp Sci Technol. wjcst.2026.f1203. https://doi.org/10.5281/zenodo.20963341